Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A method comprises: generating, by a processing entity of a computing system, a plurality of parity blocks from a plurality of lines of data blocks, wherein a first number of parity blocks of the plurality of parity blocks is generated from a first line of data blocks of the plurality of lines of data blocks; storing, by the processing entity, the plurality of lines of data blocks in data sections of memory of a cluster of computing devices of the computing system by distributing storage of individual data blocks of the plurality of lines of data blocks among unique data sections of the cluster of computing devices in accordance with a read/write balancing pattern; and storing, by the processing entity, the plurality of parity blocks in parity sections of memory of the cluster of computing devices by distributing storage of parity blocks of the plurality of parity blocks among unique parity sections of the cluster of computing devices in accordance with the read/write balancing pattern.
2. The method of claim 1 , wherein storing the plurality of lines of data blocks and the plurality of parity blocks in the data sections and the parity sections, respectively, is further in accordance with a restricted file system that operates via: a logical address mapping for a table that includes a plurality of partitions, wherein each partition of the plurality of partitions includes a plurality of segment groups, wherein each segment group of the plurality of segment groups includes a cluster number of segments, wherein each segment of the cluster number of segments includes a corresponding plurality of the individual data blocks, and wherein the logical address mapping stores the table in logical address space of the memory of the cluster of computing devices in order of the plurality of partitions, the plurality of segment groups, the cluster number of segments, and the corresponding plurality of the individual data blocks.
This invention relates to a data storage system that organizes data blocks and parity blocks in a structured manner within a restricted file system. The system addresses the challenge of efficiently managing and retrieving data in distributed storage environments, particularly in clusters of computing devices. The file system uses a logical address mapping to organize data into a hierarchical structure comprising partitions, segment groups, and segments. Each partition contains multiple segment groups, and each segment group consists of a fixed number of segments. Each segment, in turn, holds a predefined number of individual data blocks. The logical address mapping ensures that the data is stored and accessed in a consistent order, first by partitions, then by segment groups within each partition, followed by segments within each segment group, and finally by individual data blocks within each segment. This hierarchical organization simplifies data management, improves access efficiency, and ensures data integrity through the use of parity blocks stored in designated parity sections. The system is designed to operate in a cluster of computing devices, where the logical address space is shared and managed collectively.
3. The method of claim 1 , wherein the cluster of computing devices includes a number of computing devices that equals a number of the individual data blocks in a line of data blocks of the plurality of lines of data blocks plus a number of parity blocks created from the line of data blocks; wherein each computing device in the number of computing devices includes a unique data section of the data sections for storing an individual data block of individual data blocks corresponding to a segment of a segment group in accordance with a partition of a table; wherein each computing device in the number of computing devices includes a unique parity section of the parity sections for storing one or more parity blocks corresponding to a cluster number of lines of data blocks corresponding to the segment of the segment group in accordance with the partition of the table.
This invention relates to distributed data storage systems, specifically addressing the challenge of efficiently storing and managing data across a cluster of computing devices while ensuring data redundancy and fault tolerance. The system organizes data into a structured format, dividing it into lines of data blocks, where each line includes both data blocks and parity blocks generated from those data blocks. The cluster of computing devices is sized to match the total number of data blocks and parity blocks in a line, ensuring each computing device is responsible for storing a unique segment of data and a corresponding parity section. Each computing device holds a distinct data section for storing an individual data block from a segment of a partitioned table, along with a unique parity section for storing one or more parity blocks. These parity blocks are derived from multiple lines of data blocks corresponding to the segment, enhancing data reliability. The system distributes both data and parity blocks across the cluster, allowing for efficient data reconstruction in case of device failure. This approach optimizes storage utilization and ensures data integrity by leveraging distributed redundancy.
4. The method of claim 3 , wherein storing the plurality of lines of data blocks and the plurality of parity blocks in the data sections and the parity sections, respectively, further includes applying, by the processing entity, a mathematical function to fill voids in the unique data sections, wherein the mathematical function includes a logical address adjustment that effectively pushes up individual data blocks in the unique data sections to fill the voids, wherein a void of the voids corresponds to a parity position of a code line associated with a line of data blocks of the plurality of lines of data blocks.
This invention relates to data storage systems, specifically methods for efficiently storing data and parity blocks in a distributed manner to improve storage utilization and reliability. The problem addressed is the presence of voids or gaps in data sections when storing data blocks alongside parity blocks, which can lead to inefficient use of storage space and potential data integrity issues. The method involves storing multiple lines of data blocks and corresponding parity blocks in separate data and parity sections of a storage system. To optimize storage, a mathematical function is applied to fill voids in the data sections. This function includes a logical address adjustment that effectively shifts individual data blocks upward within the data sections, eliminating gaps. The voids being filled correspond to parity positions in a code line associated with a line of data blocks. By adjusting the logical addresses, the method ensures that data blocks are packed tightly, reducing wasted space and maintaining data integrity. The parity blocks are stored in dedicated parity sections, allowing for efficient error detection and correction. This approach enhances storage efficiency while preserving the ability to recover data in case of failures.
5. The method of claim 3 , storing the plurality of lines of data blocks and the plurality of parity blocks in the data sections and the parity sections, respectively, further includes applying, by the processing entity, a mathematical function to fill voids in the unique data sections, wherein the mathematical function includes a logical address adjustment that effectively pushes down individual data blocks in the unique data sections to fill the voids.
This invention relates to data storage systems, specifically methods for efficiently managing data blocks and parity blocks in storage devices to optimize space utilization. The problem addressed is the presence of voids or gaps in data sections due to deleted or relocated data blocks, which can lead to inefficient storage usage and reduced performance. The method involves storing data blocks and parity blocks in separate data and parity sections of a storage medium. To handle voids in the data sections, a mathematical function is applied to adjust logical addresses of the data blocks. This function effectively "pushes down" individual data blocks within the data sections, filling the voids by reordering the blocks without physically moving them. The parity blocks remain stored in their respective parity sections, ensuring data integrity while improving storage efficiency. The mathematical function used for logical address adjustment may include operations such as arithmetic or bitwise manipulations to recalculate the logical addresses of the data blocks. This approach allows the storage system to maintain contiguous data sections by logically reassigning block positions, thereby eliminating gaps and optimizing storage capacity. The method is particularly useful in systems where physical block relocation is costly or impractical, such as in solid-state drives or distributed storage environments.
6. The method of claim 3 , wherein storing the plurality of lines of data blocks and the plurality of parity blocks in the data sections and the parity sections, respectively, further includes applying, by the processing entity, a mathematical function to fill voids in the unique data sections, wherein the mathematical function includes using individual data blocks from every “n” lines of data blocks, using data blocks of “n-d” lines of the n lines of data blocks to fill the voids in “n-k” lines of data blocks in the “n” lines of data blocks, wherein “n” equals the number of computing devices in the cluster of computing devices, “k” equals the number of parity blocks created per line of data blocks, and “d” equals the number of the individual data blocks in the line of data blocks.
This invention relates to distributed data storage systems, specifically methods for efficiently storing and recovering data in a cluster of computing devices. The problem addressed is the need to optimize storage space and ensure data redundancy in distributed systems where data is divided into blocks and stored across multiple devices, with parity blocks providing fault tolerance. The method involves storing data blocks and parity blocks in separate sections of the storage system. To handle voids (unused spaces) in the data sections, a mathematical function is applied. This function uses individual data blocks from every "n" lines of data blocks, where "n" is the number of computing devices in the cluster. The function then selects data blocks from "n-d" lines of the "n" lines to fill the voids in "n-k" lines of data blocks. Here, "k" is the number of parity blocks created per line of data blocks, and "d" is the number of individual data blocks in a line. By strategically redistributing data blocks, the method ensures efficient use of storage space while maintaining data integrity and redundancy. This approach minimizes wasted storage capacity and improves fault tolerance in distributed storage environments.
7. The method of claim 1 further comprises: generating, by the processing entity, a second number of parity blocks of the plurality of parity blocks from a second line of data blocks of the plurality of lines of data blocks; storing, by the processing entity, the second line of data blocks in the data sections of memory of the cluster of computing devices of the computing system in accordance with the read/write balancing pattern; and storing, by the processing entity, the second number of parity blocks in the parity sections of memory of the cluster of computing devices.
This invention relates to distributed storage systems, specifically methods for improving data redundancy and reliability in a cluster of computing devices. The problem addressed is ensuring data integrity and availability in large-scale storage systems where hardware failures can lead to data loss. The solution involves a method for distributing data and parity blocks across multiple computing devices in a balanced manner to optimize read and write operations while maintaining fault tolerance. The method includes generating a first set of parity blocks from a first line of data blocks and storing these data blocks in memory sections of the cluster according to a predefined read/write balancing pattern. The parity blocks are stored in designated parity sections of the memory. The method further extends to generating a second set of parity blocks from a second line of data blocks, storing these data blocks in the data sections of the cluster's memory following the same balancing pattern, and storing the second set of parity blocks in the parity sections. This approach ensures that data and parity blocks are distributed across the cluster to minimize the impact of device failures and maintain efficient access patterns. The balancing pattern optimizes performance by distributing read and write operations evenly across the cluster, reducing bottlenecks and improving overall system reliability.
8. The method of claim 1 , wherein the data sections of memory of the cluster of computing devices each comprises: a plurality of segment group data sections for storing corresponding data segments of a plurality of segment groups.
A method for managing data storage in a distributed computing system addresses the challenge of efficiently organizing and accessing data across multiple computing devices in a cluster. The system divides memory into distinct data sections, each containing multiple segment group data sections. These segment group data sections store corresponding data segments from different segment groups, allowing for structured and scalable data distribution. Each segment group represents a logical division of data, enabling parallel processing and improved fault tolerance. The method ensures that data segments are distributed across the cluster, optimizing storage utilization and access performance. By organizing memory into segment group data sections, the system facilitates efficient data retrieval and management, supporting large-scale distributed applications. The approach enhances data locality and reduces latency by minimizing cross-node communication during data access operations. This method is particularly useful in high-performance computing environments where data distribution and retrieval efficiency are critical.
9. The method of claim 1 , wherein the parity sections of memory of the cluster of computing devices each comprises: a plurality of segment group parity sections for storing corresponding parity segments of a plurality of segment groups.
This invention relates to distributed storage systems, specifically improving data redundancy and fault tolerance in a cluster of computing devices. The problem addressed is ensuring data integrity and availability when individual devices or storage segments fail, by efficiently distributing and managing parity data across the cluster. The system organizes memory in the cluster into parity sections, each containing multiple segment group parity sections. These sections store parity segments corresponding to different segment groups, allowing for distributed error correction. Each segment group consists of multiple data segments and their associated parity segments, which are spread across the cluster to protect against failures. The parity sections are structured to enable efficient reconstruction of lost or corrupted data by leveraging the distributed parity segments. The invention improves upon prior art by enhancing the granularity of parity distribution, reducing the risk of data loss from localized failures. By storing parity segments in a segmented manner within each parity section, the system ensures that even if a single parity section fails, other segments remain available for recovery. This approach optimizes storage efficiency while maintaining high reliability, making it suitable for large-scale distributed storage environments.
10. The method of claim 1 , wherein the processing entity comprises one or more of: one or more processing core resources of a computing device of the cluster of computing devices; one or more nodes of the computing device; one or more processing core resources of another computing device of the cluster of computing devices; and one or more nodes of the other computing device.
This invention relates to distributed computing systems, specifically methods for managing processing resources within a cluster of computing devices. The problem addressed is the efficient allocation and utilization of processing resources across a cluster to optimize performance and resource management. The invention describes a method for dynamically assigning processing tasks to various resources within the cluster, including individual processing cores, nodes within a computing device, or resources from other devices in the cluster. The method allows for flexible resource allocation, enabling tasks to be distributed across different processing cores or nodes, either within the same device or across multiple devices in the cluster. This approach improves load balancing, reduces bottlenecks, and enhances overall system efficiency by leveraging available resources dynamically. The invention ensures that processing tasks are assigned to the most suitable resources based on availability, capacity, and current workload, thereby optimizing computational performance and resource utilization in distributed computing environments.
11. A computer readable memory comprises: a first memory element that stores operational instructions that, when executed by a processing entity of a computing system, causes the processing entity to: generate a plurality of parity blocks from a plurality of lines of data blocks, wherein a first number of parity blocks of the plurality of parity blocks is generated from a first line of data blocks of the plurality of lines of data blocks; a second memory element that stores operational instructions that, when executed by the processing entity, causes the processing entity to: store the plurality of lines of data blocks in data sections of memory of a cluster of computing devices of the computing system by distributing storage of individual data blocks of the plurality of lines of data blocks among unique data sections of the cluster of computing devices in accordance with a read/write balancing pattern; and a third memory element that stores operational instructions that, when executed by the processing entity, causes the processing entity to: store the plurality of parity blocks in parity sections of memory of the cluster of computing devices by distributing storage of parity blocks of the plurality of parity blocks among unique parity sections of the cluster of computing devices in accordance with the read/write balancing pattern.
This invention relates to distributed data storage systems, specifically methods for improving data redundancy and storage efficiency in a cluster of computing devices. The system addresses the challenge of ensuring data integrity and availability while optimizing storage distribution across multiple nodes. The invention involves generating parity blocks from multiple lines of data blocks, where each line of data blocks contributes to a subset of parity blocks. These parity blocks are used for error detection and correction in the event of data loss or corruption. The system stores data blocks and parity blocks across a cluster of computing devices, distributing them among unique memory sections in accordance with a read/write balancing pattern. This distribution ensures that data and parity blocks are evenly spread across the cluster, preventing bottlenecks and improving fault tolerance. The read/write balancing pattern dynamically adjusts storage allocation to maintain performance and reliability. By separating data and parity storage sections, the system enhances parallel access and reduces contention. The invention ensures that data redundancy is maintained while optimizing storage utilization and access efficiency in distributed computing environments.
12. The computer readable memory of claim 11 , wherein storing the plurality of lines of data blocks and the plurality of parity blocks in the data sections and the parity sections, respectively, is further in accordance with a restricted file system that operates via: a logical address mapping for a table that includes a plurality of partitions, wherein each partition of the plurality of partitions includes a plurality of segment groups, wherein each segment group of the plurality of segment groups includes a cluster number of segments, wherein each segment of the cluster number of segments includes a corresponding plurality of data blocks, and wherein the logical address mapping stores the table in logical address space of the memory of the cluster of computing devices in order of the plurality of partitions, the plurality of segment groups, the cluster number of segments, and the corresponding plurality of data blocks.
This invention relates to a data storage system using a restricted file system with a specific logical address mapping structure. The system addresses the challenge of efficiently organizing and accessing data in a distributed computing environment, particularly in clusters of computing devices. The restricted file system employs a hierarchical logical address mapping for a table that includes multiple partitions. Each partition contains multiple segment groups, and each segment group consists of a fixed number of segments. Each segment holds a corresponding set of data blocks. The logical address mapping stores this table in the memory of the cluster in a structured order: partitions first, followed by segment groups, then segments, and finally data blocks. This hierarchical organization ensures efficient data retrieval and management, particularly in systems requiring high availability and fault tolerance. The system also includes parity blocks stored in designated parity sections to support data redundancy and error correction. The structured mapping allows for scalable and organized data storage, improving performance and reliability in distributed computing environments.
13. The computer readable memory of claim 11 , wherein the cluster of computing devices includes a number of computing devices that equals a number of the individual data blocks in a line of data blocks of the plurality of lines of data blocks plus a number of parity blocks created from the line of data blocks; wherein each computing device in the number of computing devices includes a unique data section of the data sections for storing an individual data block of individual data blocks corresponding to a segment of a segment group in accordance with a partition of a table; wherein each computing device in the number of computing devices includes a unique parity section of the parity sections for storing one or more parity blocks corresponding to a cluster number of lines of data blocks corresponding to the segment of the segment group in accordance with the partition of the table.
This invention relates to distributed data storage systems, specifically optimizing storage and retrieval of data across a cluster of computing devices. The problem addressed is efficient data distribution and redundancy in large-scale storage systems to ensure fault tolerance and performance. The system involves a cluster of computing devices configured to store data blocks and parity blocks. The number of computing devices in the cluster is determined by the number of data blocks in a line of data blocks plus the number of parity blocks generated from that line. Each computing device in the cluster is assigned a unique data section for storing an individual data block corresponding to a segment of a table partition. Additionally, each computing device has a unique parity section for storing one or more parity blocks. These parity blocks are generated based on a cluster number of lines of data blocks corresponding to the segment of the table partition. The system ensures that data and parity blocks are distributed across the cluster in a way that maintains redundancy and allows for efficient reconstruction of data in case of device failures. The partition of the table defines how data is segmented and distributed, ensuring that each computing device handles a specific portion of the data while contributing to the overall redundancy scheme. This approach improves fault tolerance and data availability in distributed storage environments.
14. The computer readable memory of claim 13 , further comprising: a fourth memory element stores operational instructions that, when executed by the processing entity, causes the processing entity to: apply a mathematical function to fill voids in the unique data sections, wherein the mathematical function includes a logical address adjustment that effectively pushes up individual data blocks in the unique data sections to fill the voids, wherein a void of the voids corresponds to a parity position of a code line associated with a line of data blocks of the plurality of lines of data blocks.
The invention relates to data storage systems, specifically addressing the problem of efficiently managing and reconstructing data in storage arrays, particularly when dealing with voids or missing data sections. The system involves a computer-readable memory storing operational instructions for a processing entity to handle data blocks organized in lines, where each line includes unique data sections and parity positions. The key innovation is a method to fill voids in these unique data sections using a mathematical function that adjusts logical addresses to shift individual data blocks upward, effectively eliminating gaps. This adjustment ensures that the voids, which correspond to parity positions in the code lines, are filled without disrupting the integrity of the stored data. The process involves analyzing the structure of the data lines, identifying voids, and applying the mathematical function to reorder the data blocks, thereby optimizing storage efficiency and data reconstruction capabilities. The solution is particularly useful in systems where data integrity and quick recovery are critical, such as RAID configurations or distributed storage environments. The method ensures that parity information remains accurate and accessible, even when portions of the data are missing or corrupted.
15. The computer readable memory of claim 13 , further comprising: a fourth memory element that stores operational instructions that, when executed by the processing entity, causes the processing entity to: apply a mathematical function to fill voids in the unique data sections, wherein the mathematical function includes a logical address adjustment that effectively pushes down individual data blocks in the unique data sections to fill the voids.
This invention relates to data storage systems, specifically addressing the problem of managing voids or gaps in stored data sections to optimize storage efficiency. The system involves a computer-readable memory that includes multiple memory elements storing operational instructions for a processing entity. One memory element stores instructions to identify unique data sections within a storage medium, where these sections contain data blocks separated by voids or gaps. Another memory element stores instructions to apply a mathematical function to fill these voids. The mathematical function includes a logical address adjustment that effectively shifts individual data blocks downward within the unique data sections, eliminating the gaps and compacting the data. This process improves storage utilization by reducing fragmentation and ensuring contiguous data placement. The system may also include additional memory elements for other operations, such as identifying the unique data sections or determining the logical addresses of the data blocks. The overall approach enhances data storage efficiency by dynamically adjusting data placement to minimize wasted space.
16. The computer readable memory of claim 13 , further comprising: a fourth memory element that stores operational instructions that, when executed by the processing entity, causes the processing entity to: apply a mathematical function to fill voids in the unique data sections, wherein the mathematical function includes using individual data blocks from every “n” lines of data blocks, using data blocks of “n-d” lines of the n lines of data blocks to fill the voids in “n-k” lines of data blocks in the “n” lines of data blocks, wherein “n” equals the number of computing devices in the cluster of computing devices, “k” equals the number of parity blocks created per line of data blocks, and “d” equals the number of the individual data blocks in the line of data blocks.
This invention relates to data storage and recovery in a distributed computing environment, specifically addressing the challenge of reconstructing missing or corrupted data blocks in a cluster of computing devices. The system involves a cluster of computing devices where data is distributed across multiple nodes, and parity blocks are generated to enable data recovery in case of failures. The invention focuses on a method for filling voids (missing or corrupted data sections) in unique data sections by applying a mathematical function that leverages individual data blocks from every "n" lines of data blocks. The function uses data blocks from "n-d" lines of the "n" lines to fill voids in "n-k" lines, where "n" represents the number of computing devices in the cluster, "k" is the number of parity blocks created per line of data blocks, and "d" is the number of individual data blocks in a line. This approach ensures data integrity and availability by efficiently reconstructing missing data using distributed parity and redundancy mechanisms. The system is designed to handle failures in the cluster while maintaining data consistency and minimizing recovery time.
17. The computer readable memory of claim 11 , wherein generating the plurality of parity blocks from a plurality of lines of data blocks includes applying a modulo pattern to select a position of each one of the plurality of parity blocks with respect to a corresponding one of the plurality of lines of data blocks, and wherein distributing storage of parity blocks of the plurality of parity blocks among unique parity sections of the cluster of computing devices is based on the position of each one of the plurality of parity blocks with respect to the corresponding one of the plurality of lines of data blocks.
This invention relates to distributed storage systems, specifically methods for generating and distributing parity blocks to enhance data redundancy and fault tolerance. The problem addressed is the efficient and reliable storage of parity information across a cluster of computing devices to protect against data loss in the event of device failures. The system generates multiple parity blocks from lines of data blocks using a modulo pattern to determine the position of each parity block relative to its corresponding data line. This positioning ensures that parity blocks are distributed across unique parity sections within the cluster. By distributing parity blocks based on their calculated positions, the system avoids concentration of parity data in a single location, reducing the risk of data loss if a particular section fails. The modulo-based approach allows for predictable and balanced distribution of parity information, improving fault tolerance and recovery capabilities. The method involves selecting a position for each parity block using a modulo operation, which maps each parity block to a specific section of the cluster. This ensures that parity blocks are spread across different sections, minimizing the impact of localized failures. The system dynamically adjusts the distribution as new data is added or existing data is modified, maintaining consistent redundancy across the cluster. This approach enhances reliability in distributed storage environments by ensuring that parity information is evenly distributed and accessible for data recovery.
18. The computer readable memory of claim 11 , wherein the data sections of memory of the cluster of computing devices each comprises: a plurality of segment group data sections for storing corresponding data segments of a plurality of segment groups.
A system for distributed data storage and management in a cluster of computing devices addresses the challenge of efficiently organizing and accessing large datasets across multiple nodes. The system divides data into segments and groups these segments into segment groups, which are then distributed across the cluster. Each computing device in the cluster includes memory with data sections specifically allocated for storing these segment groups. Within each data section, multiple segment group data sections are provided, each dedicated to storing corresponding data segments of different segment groups. This structure allows for parallel processing and retrieval of data segments, improving performance and scalability in distributed computing environments. The system ensures that data segments are logically grouped and physically distributed, enabling efficient load balancing and fault tolerance. By organizing data into segment groups and storing them in dedicated sections, the system optimizes storage utilization and access patterns, reducing latency and enhancing overall system reliability. This approach is particularly useful in high-performance computing, cloud storage, and big data applications where large-scale data management is critical.
19. The computer readable memory of claim 11 , wherein the parity sections of memory of the cluster of computing devices each comprises: a plurality of segment group parity sections for storing corresponding parity segments of a plurality of segment groups.
This invention relates to distributed storage systems, specifically improving data redundancy and fault tolerance in a cluster of computing devices. The problem addressed is ensuring data integrity and availability when individual storage segments or devices fail, particularly in large-scale distributed systems where traditional parity schemes may not efficiently handle segmented data across multiple nodes. The system involves a cluster of computing devices where memory is divided into parity sections. Each parity section is further divided into multiple segment group parity sections, each storing parity segments corresponding to different segment groups. This hierarchical parity structure allows for efficient error detection and correction by associating parity data with specific segments rather than entire blocks or devices. The segmentation ensures that failures in one segment group do not compromise the entire parity section, improving fault isolation and recovery efficiency. The system dynamically manages these parity sections to maintain consistency across the cluster, even as data is distributed and accessed by multiple nodes. This approach enhances reliability in distributed storage environments by providing granular parity coverage, reducing the impact of localized failures, and optimizing storage overhead.
20. The computer readable memory of claim 11 , wherein the processing entity comprises one or more of: one or more processing core resources of a computing device of the cluster of computing devices; one or more nodes of the computing device; one or more processing core resources of another computing device of the cluster of computing devices; and one or more nodes of the other computing device.
This invention relates to distributed computing systems, specifically optimizing resource allocation in a cluster of computing devices. The problem addressed is inefficient utilization of processing resources across a cluster, leading to bottlenecks and suboptimal performance. The solution involves dynamically assigning processing tasks to different processing entities within the cluster, including individual processing cores, nodes, or entire computing devices, to balance workload and improve efficiency. The system includes a computer-readable memory storing instructions for managing task distribution. The processing entity can be configured as one or more processing core resources of a computing device within the cluster, one or more nodes of the same computing device, one or more processing core resources of another computing device in the cluster, or one or more nodes of that other computing device. This flexibility allows the system to adapt to varying workloads by redistributing tasks across different levels of the cluster hierarchy, from individual cores to entire nodes or devices, ensuring optimal resource usage. The approach enhances scalability and performance by dynamically reallocating processing tasks based on real-time demand and availability.
Unknown
December 15, 2020
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.